Open source software (OSS) represents a critical form of global digital infrastructure, and yet its contributor base is worryingly homogeneous. OSS is notorious for its gender representation problem – only 10% of OSS contributors are women, compared to 27% at the broader industry level. Furthermore, OSS contributors are skewing older, as fewer younger contributors are getting involved. While the OSS literature documents the experiences of (cisgender) women well, a more holistic understanding of OSS participation is lacking, as are the perspectives of other underrepresented groups. My proposal aims to develop a deeper understanding of what enables sustained participation for underrepresented contributors, to develop a more varied typology of OSS contributions, to understand how underrepresented newcomers evaluate and assess risk in OSS projects, and to develop training that helps mitigate perceived risks and barriers to OSS entry.